Research And Implementation On Multi-type Format-preserving Encryption Techniques

In the age of big data, massive data is the foundation of data mining.Specifically, large number of data, produced by enterprise, government and individual, will be used for data mining directly or indirectly. Under this situation, how to protect the privacy information of publisher who releases data is increasingly becoming an important challenge. The traditional method to protect privacy is encryption, however, it destroys the data's attribute information so that the processed data cannot be used for data mining. Besides, the existing FPE encryption method is only limited to the basic data types, while for the data with specific data characteristics, such as identity card, telephone number and Chinese name, the efficiency and security are low and the cipher text also does not meet the requirements of FPE. Given this, we study the technologies of Format-Preserving encryption in this thesis.First of all, this thesis reviews the development process of FPE and summarizes the development situation in academy and industry. Secondly,several typical algorithms and models are analyzed and implemented in this thesis. After that, according to different application scenarios, the appropriate encryption and decryption algorithms are selected. Then,based on the FPE algorithm of the basic data types and combined with the characteristics of various data, we put forward the model of segmented encryption to achieve Format-Preserving encryption for characteristic data. In addition, considering that there is no substantial progress in the FPE of Chinese name, we present the Cycle-Prefix algorithm and encryption model based on the name database to make the cipher text conform to the naming rules. In the process of data publishing, we implement the encryption system based on the data file in order to facilitate the data encryption. It greatly improves the efficiency of data encryption. At the same time, we also implement the data encryption and decryption system in the Hadoop platform, which is really convenient for data encryption or decryption in the case of large amount of data.In the end of this thesis, the function and performance of the various modules of the encryption system are verified. We compare it with traditional FPE algorithms and analyze the advantages and improvements of this system. In summary, this thesis provides new ideas for FPE in many aspects. We aim to extend the usage scope of FPE and solve theproblem of Chinese FPE under the circumstances of big data.